CLOUD REMOVAL FROM SATELLITE IMAGES

CLOUD REMOVAL FROM SATELLITE IMAGES

Shrishti Ravikant Yadav, Darshana Sandeep Ule, Anagha Deven Shinde, Divya Anil Shinde, Prof. Srushti Jadhav

Department of Computer Engineering

Vasantdada Patil Pratisthan’s College of Engineering & Visual Arts

(University of Mumbai)                                                                                                                                                               

Mumbai, India

This paper demonstrates the studies on cloud masking algorithm. more than one strategies had been advanced to cope with the trouble of cloud detection in satellite imagery and a number of cloud protecting algorithms have been advanced for optical sensors. The primary idea of MCM algorithm is to calculate the distinction among reflectance values between clear pixels and cloud shadow infected pixels. several bands of satellite image which have large distinction values are decided on for developing Multi - temporal Cloud protecting (MCM). that is applied on snap shots from Landsat 8. The qualitative and quantitative assessment show that MCM algorithm can as it should be distinguish between cloud and cloud shadows the usage of confusion matrix. some other technique is the use of pix2pix Generative adversial networks that during assessment to different algorithms gives greater accurate outcomes through distinguishing among cloud, haze, fog which can be dealt with as noise. It makes use of log loss errors function to teach the GANS model. This paper further shows a comparative study that specifies our objective of selecting the GANS version.

Key Words:  cloudy image , dataset, pix2pix, GAN algorithm , quality image.